节点文献
运动模糊图像恢复过程中的几个关键问题
Questions in Restoration of Motion-Blurred Image
【摘要】 由于相机高速运动会导致图像模糊,在相机匀速运动假设的前提下,本文以维纳滤波算法为基础,给出了运动模糊图像的恢复模型。在用维纳滤波进行模糊图像恢复时,滤波参数λ会对振铃效应、鬼影效应以及噪声三者之间产生影响:当参数λ<0.0001时出现振铃效应,当>0.01时会出现鬼影效应;增大λ能减弱振铃效应和抑制噪声,但恢复图像的分辨率会降低。由于振铃效应和鬼影效应的存在,导致模糊图片恢复质量变差。图像的非周期性、点扩散函数的离散化以及A/D采样时x和y方向的比例非1∶1等都是产生振铃效应和鬼影效应的因素,文中也对其做了详细分析。本文对振铃效应和鬼影效应的所有分析方法,不仅适用维纳滤波算法,也同样适用于其它图像恢复算法。
【Abstract】 High speed motion of camera causes blur image. On the assumption of uniform motion of the camera , on the basis of Wiener filter algorithm, restoration model is presented on motion-blurred image. In restoring image with wiener filter, the filter parameter λ influences on ring, ghost and noise. When λ<0.000 1, ring appears, and λ>0.01, ghost appears. Increasing λ weakens ring and restrains noise, but decreases restoration image resolution. Ring and ghost cause poor image restoration from the results of non-periodicity of image, discretization of point spread function (PSF) and nonuniform property of A/D sampling in directions of x and y. All these are analyzed. The methods are suit to wiener filter algorithm, and also to other image restoration algorithm.
【Key words】 Wiener filter; motion blur; blurred image restoration, point spread function(PSF);
- 【文献出处】 电子器件 ,Journal of Electron Devices , 编辑部邮箱 ,2005年03期
- 【分类号】TP274.2
- 【被引频次】17
- 【下载频次】927